New Statistics Workshop Series on Power Analysis this Fall at LiRI
Join the LiRI-UZH Statistics Workshop Series This Fall!
The Linguistic Research Infrastructure (LiRI) at the University of Zurich (UZH) is excited to announce a free series of statistics workshops designed to enhance your research skills this fall semester! Whether you're a seasoned researcher or a graduate student just starting, these workshops will provide invaluable insights into statistical power analysis, particularly as applied in linguistic research.
Workshop Overview:
1. Power Analyses - Part I: Introduction
- Date: Wednesday, September 4, 2024, 14:15-16:00
- Location: Online via Zoom
- Speaker: Audrey Bürki
- Details: In this workshop, we will introduce the notions of statistical power, type I and Type II errors. We will discuss the consequences of insufficient power as well as ways of increasing power. We will learn to estimate power for the one-sample t-test and linear regression using simulations.
- Registration Deadline: September 2, 2024 via email to Sandra Schwab
2. Power Analyses - Part II: Power for Linear Mixed-Effects Models
- Date: Wednesday, September 25, 2024, 14:15-16:00
- Location: Online via Zoom
- Speaker: Audrey Bürki
- Details: In linguistic research, data are often analysed with mixed-effects models. This is because the the data entail repeated measurements: the same participant responds to several items in one or more conditions. Similarly, a given item may be seen by more than one participant. If the data of a study are analysed with a mixed-model, power analyses for that study must be conducted with the same mixed-effects model. Unlike for the t-test, power for a mixed-effects model cannot be computed. It can only be estimated with simulations. In this second workshop, we will learn to simulate data for a continuous dependent variable using linear mixed-effects model. This will allow us to estimate power as well as the sample size(s) needed to achieve a certain power.
- Registration Deadline: September 21, 2024 via email to Sandra Schwab
3. Power Analyses - Part III: Power for Generalised Linear (Mixed) Models
- Date: Wednesday, October 9, 2024, 14:15-16:00
- Location: Hybrid mode (Online via Zoom and in-person at UZH, Andreastrasse 15, 8050 Zurich; specific room to be announced)
- Speaker: Andri Manser
- Details: Data that is not normally distributed (e.g. proportion, count, ordinal data) is ubiquitous in language research. In this third workshop on power analyses, we will learn how to a) simulate data sets that are not normally distributed, either from scratch or based on existing pilot data, and b) how to run power analyses based on the simulated data using generalised linear (mixed) models. For clarity and easy of understanding, we will first learn to implement this "by hand" in R. We will then
- Registration Deadline: October 5, 2024 via email to Sandra Schwab
Why Attend?
These workshops offer a rare opportunity to gain hands-on experience with statistical power analysis in a collaborative learning environment. By attending, you will:
• Enhance your understanding of key statistical concepts.
• Learn to conduct power analyses using simulations.
• Gain practical skills in R, a powerful statistical software.
• Network with peers and experts in the field of linguistic research.